THE LIFE HISTORY AND TROPHIC RELATIONSHIP OFTHE NINESPINE STICKLEBACK, PUNGITIUS PUNGITIUS, IN

THE APOSTLE ISLANDS AREA OF LAKE SUPERIOR 1

BERNARD L. GRISWOLD2 AND LLOYD L. SMITH, JR.3

ABSTRACT

The ninespine stickleback is an important food of juvenile lake trout in the Apostle Islandsarea. It is the most numerous fish of the area and is distributed in deep waters during thewinter and in shallow waters during the summer. Females grow fastel than males. reachingan average total length of 80 mm at age 5. Males live to age 3 and attain an average lengthof hh mm. Annulus formation on otoliths is complete by mid-July. Seasonal growth is morethan half complete by early August: growth of mature females ·is delayed until afterspawning.

Both sexes mature over a period of 3 yr. Spawning occurs from mid-June to late July.Males apparently do not live as long as females. possibly because of a post-spawning mor­tality. Egg number is a linear function of fish length. although this relationship is differentfor fish from two separate areas. Environmental differences between these areas. which mayaffect spawning time. possibly cause the differences in fecundity. Significant quantities ofmaturing eggs atrophy just prior to spawning. a phenomenon which changes the fecundityrelationship.

Sticklebacks eat a variety of invertebrates. particularly the crustaceans Mrsis re!icra andPontoporeia alfinis. Food eaten by the stickleback and slimy sculpin is similar. bill theadaptability of both species tends to eliminate serious competition. The lake trout is theonly serious predator of stickleback.

Investigations were made by the Bureau ofCommercial Fisheries (BCF) on the biologyof the fishes of Lake Superior from the early1960's. These studies have been primarily lifehistory studies on economically valuable speciessuch as the lake trout, Sa{velil/lls l/amayclIsh(Walbaum), various coregonids, and the smelt,o.~II/('rUS morda.1" (Mitchill). Some of the morecomprehensive of these include Bailey (1964),Dryer (1963), and Dryer and Beil (1964).

When it became apparent that these fishstocks were being rapidly modified by exoticspecies, particularly the sea lamprey, Pcfl'o-

1 This work is a result of research sponsored in partby . the Minnesota Agricultural Experimental Station.Umversity of Minnesota. SI. Paul, MN 55101.

2 Northeast Fisheries Center, National Marine FisheriesService, NOAA, Woods Hole. MA 02543; present address:OhIO Cooperative Fishery Unit, Department of Zoology,OhIO State University. 1735 Neil Ave., Columbus. 01-143210.U aOepartment of Entomology, Fisheries and Wildlife.

mversity of Minnesota, SI. Paul, MN 55101.

Manuscript accepted March t973.FISHERY BULLETIN: VOL. 71, NO.4. 197.1.

myzo/I maril/lls L" additional emphasis wasplaced on ecological studies of noncommercialfishes whose life histories are interrelated withthe exploited ones. One of these, the ninespinestickleback, occurs in BCF index trawl samplesin greater numbers than any other. However,prior to the study presented in this paper, itsecological significance in these waters waslargely unknown. This study was designed toinvestigate the life history of the ninespinestickleback in the Apostle Islands region ofLake Superior and determine the importanceof its relationships to the economically impor­tant fish species of the area.

The ninespine stickleback is one of the mostwidely distributed northern fishes, occurringin freshwater, estuarine, and coastal saltwaterenvironments (Bigelow and Schroeder, 1953).It is abundant in cool-water habitats similar toLake Superior throughout the Northern Hem­iRphere. In middle western United States, it isfound in all the Great Lakes except Lake Erie

1039

FISHERY BULLETIN: VOL. 71. NO.4

:ULL

dH'GAN24

o\

N0.

vided data for stickleback age and growthstudies and supportive information about dis­tribution and relative abundance.

A small semi-balloon otter trawl with a 4.7-mheadrope, 38.1-mm mesh body (stretched meas­ure), and 12.7-mm mesh cod end was usedfrom a 16-ft aluminum boat with a 40-horse­power outboard motor at stations 2 and 3. Alarge trawl of the same type (9.4-m headrope,50.8-mm body, and 12.7-mm cod end) wasused aboard the Sis('()lI'cf for deepwater trawl­ing. Adult lake trout were Rampled with ex­perimental gill nets ranging from 1.3- to 7.6-cmmesh (bar measure). Zooplankters and stickle­backs examined for stomach contents were col­lected with a sled-borne plankton net with aI-m square mouth and 0.158-mm mesh, thebottom edge of which ran approximately 30 cmabove the lake bottom. Temperature profiles

FIGURE I.-Sampling stations in the Apostle Islands.Blacked out area in inset shows location of these islandsin Lake Superior.

The data for the present study were collectedin 1967-69 in cooperation with the BCF GreatLakes Biological Laboratory, Ann Arbor,Mich.'

Experimental sampling operations were con­ducted aboard the BCF research vessel Sis(,()ll'ctin conjunction with sampling by BCF in theApostle Islands, Lake Superior. The main ob­jectives of the BCF program were to enumeratejuvenile lake trout, monitor spawning succeRS,measure the impact of the lamprey on the troutpopulation, and contribute research on the biol­ogy of other species in the lake.

One of the main objectives of the ninespineRtickleback study was to defiine ecological re­lationships between the stickleback and otherfish, particularly the lake trout. With this ob­jective in mind, and within the limitations ofthe ongoing BCF program, sampling was es­sentially confined to seven stations within theApoRtle Islands (Figure 1). Five of these sta­tions (12, 24, 44, 75, and 86) were Rampledregularly as juvenile trout index trawl Rtationsand yielded data pertinent to questions ofspecies interaction. Comparatively, they sup­ported varying numbers of trout and stickle­backs, and historically they comprised theknown lake trout nursery areas within theislands. At present, only stationR 75 and 86support good populations of juvenile trout.The other two stations, 2 and 3, RUppOrt abun­dant midsummer stickleback populationH andcomprise shallow, inshore areas. These twostations were outside the regular BCF troutindex sampling routine and required specialsampling. Collections from these stationR pro-

(Hubbs and Lagler, 1958). It has also beentaken in inland lakes as far south as northernIndiana (Nelson, 1968a), and Moore and Braem(1965) found small numbers in some Lake Sup­erior tributaries. It 'is a fish of cool, quiet water(Hubbs and Lagler, 1958), and it is generallyabundant wherever it occurs in lakes or seaswithin its range.

METHODS AND MATERIALS

4 Presently the Great Lakes. Fi~hery Laboratory, U.S.Bureau of Sport Fisheries and WildlIfe.

1040

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

were obtained by bathythermograph at eachstation.

Trout index trawling aboard Si.~c()lVet wasdone for two 2-wk periods annually, one springand one fall. A total of 286 tows, each 1 milelong, were conducted during the sticklebackstudy from April 1967 through May 1970.Additional records from 322 index tows takenin all ice-free months dating back to April 1965were also included in the study of distribution.A total of 46 1/4 -mile tows were made fromJune through September, 1968 and 1969, withthe small outboard trawl, and 18 tows weremade with the sled at various times of the yearat the index stations in 1968 and 1969. Trawlsamples were separated by species, and the fishwere counted. After lake trout were measured,fish samples were put in jars of appropriatesize and preserved in 10% Formalin." Samplesfrom the plankton sled were flushed directlyinto jars and preserved in Formalin.

The main limitation of this sampling schemewas that the index trawl series was conductedat a limited number of stations representingtrout nursery areas. It is possible that the bio­logical communities and the resultant ecologicalinterrelationships in these areas may not berepresentative of the entire Apostle Islandsarea. Also, since areas of sampling were lim­ited and the efficiency of the index trawl wasunknown, it was impossible to estimate absolutebiomass of the stickleback population. Alter­nate methods of population estimation wereunfeasible because of the immensity of the pop­ulation and the area.

Additional methods are discussed in appro­priate sections throughout the paper.

DESCRIPTION OF GROWTH

Sticklebacks used for age and growth studieswere stratified by length, subsampled if catchesWere large, and frozen. Otoliths (sagittae) we~'e

removed, cleaned, dried in alcohol, cleared 111

creosote, and mounted on microscope slides inCanada balsam as described by Jones andHynes (1950). The otoliths contained a series

"Reference to trade names does n~)t imply endorsementby tbe National Marine Fisberies SerVIce, NOAA.

of transparent and opaque rings, and it wasevident that the transparent zone was laid downat the beginning of the growing season in lateJune and early July. Therefore, the annulus wasconsidered the outer edge of the opaque zone.Nomographs, based on the male and femalebody length-otolith length curves, respectively,were used to calculate body length at each an­nulus (Carlander and Smith, 1944).

Validation of the Otolith Method

Determination of age from otolith rings wasmade with little difficulty. In older fish the an­nulus on the outer margin was not as distinctas ill young fish, put usually could be identifiedat the posterior portion of the otolith. After twoindependent age determinations of 808 pairs ofotoliths, 37 pairs required reexamination be­cause the first two were not in agreement.

Otoliths were analyzed from sticklebackstaken with experimental trawls in 1967 and1968. The samples represented six year classesand five age-groups. The otoliths were from 372fish taken at station 2 and 436 taken fromstation 3. Length-frequency distributions wereestimated from 18,175 fish measured from 80trawl samples.

The following evidence indicates that annulion otoliths provide an accurate measure ofstickleback age and growth:

1. Length distribution by age-group basedon otolith readings and modes in length­frequency distribution of all measuredfish were in close agreement; h<».Veve1',

most distributions exhibited some degreeof overlap (Figure 2 and 3). In the Junesample (Figure 2), the distribution of age­group 0 fish truncated on the left-handside because it had been discovered thatall of these smaller fish were of O-group.

2. Calculated lengths at the end of each yearof life corresponded well between thevarious year classes (Tables 1 and 2).

3. There was a general increase in averagebody length with assigned age of the fish(Tables 3 and 4).

4. The average empirical length of aged fishcolleded in spring prior to annulus for-

1041

TOTAL FISH LENGTH (MM)

FISHERY BULLETIN: VOL. 71, NO.4

TABLE I.-Average calculated total lengths in millimetersof female ninespine sticklebacks from stations 2 and 3 inthe Apostle Islands.

Number Years of lifeYear of

Station class fish 2 3 4 5

3 1962 I 45.50 58.75 65.75 74.00 81.002 1963 16 46.25 58.25 66.25 73.503 1963 19 48.75 61.53 68.56 76.62 79.002 1964 24 44.90 61.20 69.163 1964 40 47.38 61.64 68.79 74.922 1965 60 48.65 61.81 70.503 1965 122 48.66 63.25 70.662 1966 110 47.33 63.223 1966 70 48.47 61. 182 1967 24 49.003 1967 2 47.75

'2 234 234 88 23 8'3 254 254 62 38 9

:!2 47.22 61.87 68.63 73.5023 47.75 61.27 68.44 75.18 80.00

Average empirical lengthof fish in spring priorto annulus formation 47.08 62.32 71.16 7753 81.00

I Total number of fish.2 Grand average calculated length.

IN - 281

( N - '2481

O~----------"",o:£1~~~-1(N - 12)

200

I~O

100

~O

a

a:x:U> I~

"-"- 100

crwOJ

":J az

FIGURE 2.- Length-frequency distribution of sticklebacks(top graph) and length frequencies of age-groups 0 to IVbased on otolith reading.~ (lower graphs) of fish collectedin June 1967. N equals number offish.

TABLE 2.-Average calculated total lengths in millimetersof male ninespine sticklebacks from stations 2 and 3 inthe Apostle Islands.

Number Years of lifeYear of

Station class fish 3

3 1963 8 45.21 60.23 66.132 1964 8 44.25 59.25 65.503 1964 38 43.97 59.732 1965 12 44.25 59.623 1965 70 47.30 60.472 1966 108 49.72 62.703 1966 44 48.54 59.702 1967 10 50.603 1967 22 47.00

'2 138 138 38 8'3 182 182 50 2

"2 47.20 60.52 65.50"3 46.40 60.03 66.13

67.6060.6048.55

I Total number of flsh." Grand average calculated length.

Average empirical lengthof fish in spring priorto annulus formationI N-~~)

IN-2431)

( N-301

o

350

300

250

200

150

'00

:x: 50~"- a"-0

crwOJ

":JZ

FIGURE 3.- Length-frequency distribution of sticklebacks(top graph) and length frequencies of age-groups 0 to IIIbased on otolith readings (lower graphs) of fish collectedin August 1967. N equals number offish.

TOTAL FISH LENGTH IMMI

30 40 50 60 70 80 mation agreed closely with grand averageback-calculated length of each assignedage-group (Tables 1 and 2). Empiricallengths slightly exceeded calculatedlengths. indicating seasonal growth mayalready have begun.

1042

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

TABLE 3.~ Length frequency of the age-groups of femalesticklehacks taken in July 1968 from the Apostle Islands.

TABLE 4.~Length frequency of the age-groups of malesticklehacks taken in July 1968 from the Apostle Islands.

558

188 14 42 14 2

66.7 69.4

54 9

Totollength(mm)

40-4142-4344-4546-4748-4950-5152-5354-5556-5758-5960-6162-6364-6566-6768-6970-7172-7374-7576-7778-7980-8182-8384-8586-87

Age-groups Totollength

II III iV V (mm)

5 46-474 48-49

8 50-5110 52-539 54-55

18 56-576 58-598 3 60-61

10 9 62-637 20 64-657 14 66-67

4 13 68-694 17 2 70-711 10 1 72-731 17 6 74-75

28 1216 16 Average

6 15 3 length

3 10 27 4 Number1 of fish

5142

1112117

13741

51.8

78

Age·groups

II III

The relation:'\hip between total fiRh lengthand otolith length waR eRtimated from otolithRof 320 males and 488 females. Body lengthsat each 1.0-mm interval of otolith length wereaveraged, and the mean total lengths were plot­ted againRt the otolith length. The relationRhipRfor fiRh from RtationR 2 and 3 were very nearlythe Rame, but the relationship for males andfemales were empirically quite different (Figure4). Second degree polynomial regressions werefitted to these data and the resulting equationsWere:

y = - 25.864 + 2.146X - 0.01lX2Y = - 25.075 + 2.131X - 0.0010X1

Growth in Length

• MALEn FEMALE

70

80

°0~-~'0:------;;2~0-~3t;0:--~40:;------;5!;::0-----,';60'-----:7'-=-0--8,L0:-----l

PROJECTED ANTERIOR OTOLITH RADIUS (MM)

! 60

J:....'"z 5"j>­o~ 4

...Jg

.... 3

:iw:E

FIGURE 4_~Body length-otolith length relationship forApostle Islands ninespine sticklehacks. Solid line. males;dashed line. females.

3

78.5

12

76.8

70

73.8

157

66.5

102

52.3

MalesFemales

Numberof fish

Averagelength

where Y = total fish length (mm) andX = projected anterior otolith

radius (mm).

CoefficientR of determination (r2 ) values were0.978 and 0.958 for male and female equationR,respectively. An analysis of covariance showedthe elevation was significantly different at the

0.01 level (F1,78 = 48.32). Therefore, growthparameters were calculated separately for eachsex. In Figure 4, the plots were extrapolated tointercept the body-length axil' at 6 mm, thelength of otolith formation of AlaRkan three­Rpine Rticklebacks (JoneR and Hynes, 1(50).

Growth rates were determined for femalesfor the 1962 through 1967 year classes (Table

1043

1) and for males for the 1963 through 1967year classes (Table 2). Females averaged 47.5,61.6, 68.5, 74.3, and 80.0 mm in total lengthat the first through fifth annuli, respectively.Age-groups I, II, and III in males averaged46.7, 60.2, and 65.6 mm. There was little dif­ference between growth rates of fish from sta­tions 2 and 3. Females from station 2 werelonger at the end of 2 of 5 yr of life, and malesfrom station 2 were longer at the end of 2 of 3yr of life. Females grew faster than males dur­ing the first year, being 0.02 and 1.35 mmlonger at stations 2 and 3, respectively, at an­nulus 1. These differences increased to almost3 mm after 3 yr. Females reached age 5, malesage 3.

Both sexes showed the greatest growth inlength in their first year (Tables 5 and 6), whenfemales attained about 60% and males about70% of maximum size. The average annual in­crement calculated from the summation ofannual increments was about 30% for bothsexes during the second year and on the orderof 10% per annum thereafter.

Annulus formation was evident in somesticklebacks as early as June in 1967, and allannulus formation was complete by the secondweek in July. Annulus formation occurred be-

FISHERY BULLETIN: VOL. 71, NO.4

fore spawning and started when surface watertemperatures reached about 10 0 e in 1967 and1968.

Percentage of annual growth completed atany date during the growing season was com­puted by comparing the length at capture on agiven date in 1967, an average growth year asindicated by average annual increments, withthe calculated length at annulus formation in1968 (Table 7). Male growth was faster at thebeginning of the growing season with about 50%of the annual growth being completed by lateJuly. Early seasonal growth of females wasslower than in males, and 50% of the annualgrowth was not complete until mid-August.Growth then declined and was complete ornearly complete in all fish by early Decemberwhen fall turnover occurred. These calculationsare consistent with length-frequency distribu­tions of fish from stations 2 and 3 (Figures 5and 6). These monthly distributions show thatgrowth, as indicated by increasing modal lengthof the two apparent age-classes, had begun byJuly and continued into December.

Older fish had a tendency to grow slower intheir first years of life than the younger fish(Tables 1 and 2). This growth pattern (Lee'sphenomenon) has been reported for many fish

TABLE 5.-Grand average increment in total length of female ninespine sticklebacks from stations 2 and 3 expressedin millimeters.

Year Percentageof Sums of of

Station life 1962 1963 1964 1965 1966 1967 Mean means increment

2 1 46.25 44.90 48.65 4733 49.00 47.22 47.223 1 45.50 48.75 47.38 48.66 48.47 47.75 47.75 47.752 2 12.00 16.30 13.16 15.89 14.33 61.55 30.33 2 13.25 12.78 14.26 14.59 12.71 13.51 61.26 28.22 3 8.00 7.96 8.69 8.21 69.76 13.33 3 7.00 7.03 7.15 7.41 7.14 68.40 11.62 4 7.25 7.25 77.01 10.33 4 8.25 8.06 6.13 7.48 75.88 10.93 5 7.00 2.38 4.69 80.57 6.1

TABLE 6.-Grand average increment in total length of male ninespine sticklebacks from stations 2 and 3expressed in millimeters.

Year Sums Percentageof of of

Station life 1963 1964 1965 1966 1967 Mean means increment

2 1 44.25 44.25 49.72 50.60 47.20 47.203 1 45.21 43.97 47.30 48.54 47.00 46.40 46.402 2 15.00 15.37 12.98 14.45 61.65 30.63 2 15.02 15.76 13.17 11.18 13.78 60.18 29.62 3 6.25 6.25 67.90 10.13 3 5.90 5.90 66.08 9.8

1044

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

TABLE 7.-Percent growth completed at various dates ofthe 1967 season for ninespine sticklebacks in the ApostleIslands.

48 75 28M ~ ~ M 54

100 100 100 100 97

Date

July 22·27August 8-19December 4

o

Male

II o

Female

II

52779

III

376085

IV

75

100

APRIL

10 (N=156)

JUNE

10 (N' 906)

80605040

TOTAL LENGTH

30

NOVEM8ER(N-258)

SEPTEMBER( N - 52)

JULY(N-792)

20

DECEMBER10 (N-361)

10

10

20

10

AUGUST(N·1768)

1°L..m.L------"'~b+o..~

longer. This tendency does not explain thepresence of "Lee's phenomenon" among fe­males, where it is most prevalent. Figures 5and 6 demonstrate that young-of-the-year fishwere being sampled throughout the latter partof each season. Therefore, it is possible thatbiased sampling of precocious individuals inthe younger age-groups creates the appearancethat younger fish of both Sexes are growingmore rapidly than older fish did at the sameage. To minimize these suspected biases in thegrowth-rate calculations, back calculations foreach year dass over successive collection yearswere averaged as unweighted means.

Growth was fairly stable from year to year(Tables 5 and 6) with the exception of 1964.No explanation for this poor growth year wasdiscovered.

FIGURE 6.- Lengt h-frequency histograms of st ation 3sticklebacks expressed in millimeters. N equals number offish in sample.

1LI...J0..:::E«(J)

...J«f­of-

LLof­Z1LIUc::1LI0..

MAY

10 (N' 76)

JUNE

10 (N-342l

w...J

JULY0..~ 10 (N- 1313)<r(J)

...J<r AUGUSTf-0 (N- 663)f- lO

LL0f-

SEPTEMBERZw 10 (N-1I15)uCl::W0..

10

NOVEMBER

10 (N-297)

20 30 40 50 60 80

APRIL10 (N- 237)

species (Lee, 1912). In the Lake Superior study,there appeared to be no significant disappear­ance of faster growing individuals from thelength-frequency distributions (Figures 5 and6). However, as will be shown below, theredoes appear to be a sex-specific mortality asso­ciated with spawning. Since faster growingmales tend to spawn at an earlier age, slowergrowing males may be left in the population

TOTAL LENGTH

FIGURE 5.- Length-frequency histograms of station 2sticklebacks expressed in millimeters. N equals number offish in sample.

1045

Ninespine sticklebacks from Lake Superiorgrow as large, or larger, than those reported inother areas. The most comprehensive growthcalculations on the ninespine stickleback priorto this study was done by Jones and Hynes(1950) on a population in a small Englishstream. They found fish reached only 37 mmin 1 yr and 46 mm in 3 yr. The largest fish re­corded was 55 mm. These fish were in directcompetition with a population of threespinesticklebacks. Nelson (1968b) found the largestfish in a northern Indiana lake to be 67.3 mm.Bertin (1925) and Leiner (1934) both recordEuropean ninespine sticklebacks over 80 mm,but Blegvad (1917) found none over 50 mmin Danish brackish waters. McKenzie andKeenleyside (1970) report fish up to 75 mm inLake Huron. The formula, TL = 1.14SL, wascalculated from 300 Lake Superior fish andused to convert reported SL (standard length)to TL (total length) for these comparisons.

Walford (1946) proposed a transformationof growth data to be expressed as a straightline described by the equation:

L t + 1 = Loo (l-K) + KL p

body length at age t + 1body length at age tultimate lengthslope.

The ultimate length is found graphicallywhere the Walford line intersects a line drawnat 45 0 through the origin of the graph and istheoretically the greatest length a fish of thepopulation will attain. The slope of the line, K,indicates the decrease in growth rate.

Walford lines were plotted for Apostle Is­lands sticklebacks from stations 2 and 3 (Figure7). Means of stations 2 and 3 grand averagecalculated lengths at each annulus were used tocombine station data for each sex.

Calculated values were:

Males K 0.411LOCi 69.690

Females K 0.672Lx 87.660

The ultimate length for males is less than ex-

1046

FISHERY BULLETIN: VOL. 71, NO.4

90

80

E 70E

+ 60 .".;,/'...", .../

w 50 ...... ./Cl ... ~~ ./<l: ~~

./I- 40<l: ./

:l: ./I- 30Clzw 20...J

10

OL----l.-.....L-_-L-_.L-_L..---..l_-I..._--L---l

o 10 20 30 40 50 60 70 80 90

LENGTH AT AGE t (mml

FIGURE 7.-Walford lines for Apostle Islands ninespinesticklebacks. Short dashed line, males; long dashed line.females.

pected from what is known about male lengthfrequency (Table 4). This is probably due tothe small number of points used because of theshort life span. More points can be includedfor the female relationship, and the ultimatelength is reasonable.

Growth in Weight

Plots of weight against length were made formales and females by calculating average weightsfor fish at I-mm intervals. Weights for femalesin the mature size range were less than maleweights. The cause for this was the abrupt dropin weight of females which had just spawned,since the mature female gonad accounts for upto 10% oftotal fish weight. After females collectedin late June and July were omitted from thelength-weight plots, the lines for males andfemales appeared similar.

Analysis of covariance showed slope, and in-tercept differences were nonsignificant at the

0.05 level (Fl,745 = 2.71 and Fl,746 = 1.36, res­pectively), and the data were pooled to calcu­late a length-weight relationship. The resultantregression included 748 fish and was describedby the formula:

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

log W = - 4.5681 + 3.2190 log L

where log W = common log of the weight (O.lg)and log L = common log of the length (mm).

Growth in weight was calculated on the basisof weight at the time of each annulus formation(Table 8). The greatest increment in weightoccurred in the second year of life for both sexesat both stations. After the second year, growthin weight declined rapidly, but the decline wasmore pronounced in males.

Instantaneous rates of growth in weight be­tween consecutive ages varied from 0.804 to0.223 for males and from 0.842 to 0.192 for fe­males.

may well be the explanation for this; however,no historical records of stickleback abundancein the Duluth area are available to help justifythis hypothesis.

Temperature profiles were secured concur­rently with trawl tows. Sticklebacks were foundin the warmer deep water in early spring, butwere evenly distributed at all depths duringspring turnover (Table 9). In midsummer theycongregated on the warm shallow snoals wherewater temperatures reached as high as 20°C.When the fall turnover occurred in September,they were again fairly evenly distributed at alldepths. By November the fish were beginning tocongregate in the deeper water and were foundthere in great numbers during December. No

TABLE S.-Calculated weight attained by Apostle Islands sticklebacks at the time of annulusformation.

Male Female

Total Calculated Total CalculatedAge· length weight Percentage length weight Percentagegroup (mm) (g) increment (mm) (g) increment

Station 2I 47.20 0.662 47.22 0.663

II 60.52 1.470 122.0 61.87 1.580 138.3III 65.50 1.821 23.9 68.63 2.215 40.2IV 73.50 2.751 24.2

Station 3I 46.60 0.627 47.75 0.687

II 60.03 1.401 123.4 61.27 1.494 117.4III 66.13 1.955 39.5 68.44 2.185 46.2IV 75.18 2.955 35.2V 80.00 3.624 22.6

DISTRIBUTION

While some species of sticklebacks can befound in a marine environment far from shore,the ninespine stickleback is an inshore species(Bigelow and Schroeder, 1953). Hubbs andLagler (1958) state that the species is abundantin the I,;le Royale and Keewenaw Bay areas ofLake Superior. BCF operations found stickle­backs in Nipigon Bay, on the north shore of thelake. While sticklebacks were found throughoutthe Apostle Islands, it should not be concludedthat these fish are similarly abundant on allLake Superior shorelines. Anderson (1969) re­covered only 0.1 stickleback per trawl tow in 90tows in the Duluth area, sampling during allice-free months. Degradation of water quality

TABLE 9.-Average numbers of sticklebacks caught per30-min trawl tow each month at various depths during1967-69.

[Numbers of tows are in parentheses.l

Depth(fathoms) Apr. May June July Aug. Sept. Nov. Dec.

3·9 0.0 2.7 12.8 615.7 641.5 75.4 0.3(2) (4) (6) (16) (12) (5) (5)

10·26 0.3 8.4 14.8 10.5 11.4 80.0 18.2( 15) ( 11) (45) ( 17) ( 16) (42) ( 15)

27·50 58.4 11.7 32.7 0.1 0.0 147.1 114.7 267.3( 12) (3) ( 16) (6) (1) (9) (11) (4)

data are available for midwinter, but it isassumed the deepwater habitat is used at thistime since they were abundant there the fol­lowing spring. Nelson (l968b) also noted sum-

1047

FISHERY BULLETIN: VOL. 71. NO.4

RELATIVE ABUNDANCE

TABLE 10.-Rank correlation coefficients (r,) betweenstickleback CPE and CPE's of other species in 100 ran­domly selected index trawl tows in the Apostle Islands.1965-68.

relationships at other times of the year orother places, nor do they consider possible ef­fects of differences in availability to the gear.Further discussion of these correlations will bepresented in the section on food habits.

Dryer (1966) reported that the ninespinestickleback was the most abundant species in578 Apostle Islands experimental trawl towsfrom 1958 through 196:3. The average CPE ofstickleback during thiR period was 51.8. Smeltranked second (CPE = 40.4), and Rlimy sculpin,Cottus cog1latus Richardson, were third (CPE= 22.3). Lake trout CPE was 6.8.

Over the 1965-68 period, 322 trout indextows were made, and the relative rank of Rtickle­backs, smelt, and slimy sculpinR was unchangedfrom the 1958-63 period (Table 11). The CPE'swere 61.1, 50.3, and 17.7, reRpectively. The catchRtatisticR of the two time periods show slightincreaRes for sticklebacks, smelt, and lake trout.Further comparison is meaningless however,because sampling allocations between indexstations used in the two time periods werenot identical, and extreme between-tow variationwas observed. Table 11 does not include datafrom stations 2 and 3 because they were notused by BCF as trout index trawl stations.

To sample the shallow inshore area of station2, where sticklebacks congregated in midsum­mer, 1,4 mile tows were made with the smalloutboard trawl. The wingspread opening of thetrawl was determined by measuring the widthof the trawl path on the bottom as indicated byground-rope disturbance immediately after the

-0.077+ 0.073-0.461"+ 0.568"-0.360"

Species

**Signiflcant at 0.01 level (98 degrees of freedom).

Juvenile lake trout to 11 inchesSmelt adultsSmelt juvenilesSlimy sculpinTrout -perch

mer inshore migrations of ninespine stickle­backs in Crooked Lake, Ind.

The shoal areas where sticklebacks concen­trated in the summer were represented bystations 2 and 3 (Figure 1). Station 2 includeda sandy beach, approximately 2 miles long,at Presque Isle Bay, Stockton Island, whichgradually deepened and merged with the softbottom muds of the deeper areas. In summersticklebacks congregated in large numbers onthe beach in Ih to 2 fathoms of water. Thefish could easily be seen by eye and wereevenly distributed over the bottom throughoutthe entire beach area. Station 3, along thesouth shore of Oak Island, was characterizedby a sandy shelf about 1/4 mile wide parallel­ing the shore which gradually deepened to 8fathoms. The bottom then dropped steeply to12 fathoms. This steep slope was covered witha dense growth of Nitella sp. The summer con­centration of sticklebacks occurred in thisvegetation. No fish could be seen over thesandy shelf, and no fish were caught unlessthe net was recovered with mats of Nitella sp.clinging to it.

Dryer (1966), Anderson and Smith (1971), andfield observations by the senior author of thispaper indicate that smelt, slimy sculpins, laketrout under 28-cm TL, and trout-perch, Percop­sis o/llis('o/llaycus (Walbaum), had food habitssimilar to the stickleback and were apparentlyabundant enough to be potential competitorswith the stickleback. Rank correlation coef­ficients (1'.) were obtained between the aver­age number of sticklebacks caught per I-miletrawl tow (CPE) and the CPE of these potentialcompetitors in 100 trout index tows. The CPE'swere ranked for each species, ties beinggiven the mean rank, and r .. values were deter­mined by the formula presented by Snedecor(1956). The 100 tows were randomly selectedfrom all (322) trout index tows made at stations86, 75, 44, 24, and 12 from 1965 to 1968. Thesignificant relationships were those for juvenilesmelt, trout-perch, and slimy sculpins (Table 10).Sticklebacks inhabited different areas thanjuvenile smelt and trout-perch, and stickle­backs and sculpins inhabited the same areas,at least during spring and fall. The results ofthe correlations do not preclude significant

1048

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

TABLE It.-Average catch per I-mile trawl tow (ePE) at various index stations in the Apostle Islands

calculated from Bureau ofCol11mercial Fisheries index trawling data. 1965 through 1968.

Station

Species tWeighted

86 75 24 12 44 mean

Number of tows 99 98 28 76 21

Stickleback 42.13 93.22 53.52 60.01 4.69 61.11Smelt 88.44 19.87 20.71 52.33 7.90 50.30Lake trout 23.16 7.63 4.29 4.63 .76 10.71Slimy sculpin 7.93 19.63 38.43 20.01 18.57 17.71Fourhorn scul pin .38 .72 .36 4.20 1.14 2.00Spoonhead sculpin .35 .47 .61 3.29 .14 1.06Mottled sculpin .04 .01 .01Lake whitefish .76 1.06 .14 12 3.62 81Bloater 4.44 5.38 7.30 .33 4.63Short jaw cisco .08 .06 18 08Kiyi .01 .01 <.01lake herring .12 .05 .14 .08Unidentified coregonid 5.80 56 1.37 1.24 230Round whitefish 1.00 13.93 13.33 2.03Pygmy whitefish .28 .71 .25 1.24 .86 .65Alewife .01 .04 .01Trout-perch 3.08 1.02 .11 .17 7.2B 1.69Burbot .17 1.03 .43 .45 .10 50long nose sucker .07 .14 06Johnny darter .01 <.01

I Taken from: Bailey, R. M. (chairman). 1970. A list of common and scientific names of fishes from the UnitedStates and Canada. Am. Fish. Soc., Spec. Pub!. 6, 149 p.

trawl waR towed through the area. ThiR allowedei"timation of minimum deni"ity.

Catches with this trawl were highly variableeven though visual Rightings, during both dayand night, indicated fiRh were quite evenlydiRtributed. Light obviouRly influenced catchbecauRe no fiRh were caught during the day­light hours even though they could be seen.Moonlight intenRity probably influenced thecatch, as did wind direction and intensity, byadverRely affecting the efficiency of the Ramp­ling operation. It is entirely pORRible theRefactorR alRo influenced catcheR by Si.~colI·ct indeeper waterR,

Summer nighttime trawl tOWR on the sandybeach, station 2, usually caught from 300 to2,500 RticklebackH in 1f4 mile (Table 12). TheRe

TABLE 12.-Catches of sticklebacks in 1967 and 1968 onthe sand beach of station 2 during the nigbt using the4.7-111 otter trawl and outboard.

MonthNumber of Total Average

tows sticklebacks per tow

April 5 6 1.2June 3 217 72.3July 10 6,985 698.5August 8 7,564 945.5September 4 3,622 905.4November 4 0 0.0

values correspond to minimum density estimatesof 0.18 to 1.49 fish per square meter. Sticklebackswere the only species caught in the area duringtheir summer concentration, but johnny darters.Etheostoillu Iliyrulll RafineRque, pygmy white­fiRh, Prosopillill co//lte/'i (Eigenmann and Eigen­mann), trout-perch, and smelt were found therein spring and fall.

REPRODUCTION

Maturation

The gonads of 401 spring-collected fish fromwhich otoliths were extracted were analyzedfor maturity. In addition, ;14 male and 72 femaleI annulus fiRh (determined from length-frequen­cy histogramR) were included in maturity studieR.

Both male and female RticklebackR maturedover :3 yr (Table 1:3). A greater percentage ofmales matured at age 1, with large Rize being aprerequiRite for early maturity. By age 2, nearlyall fish of each sex were sexually mature, butneither sex was 100% mature until the thirdyear of life. Jones and HyneR (1950) found 99%of a sample of European ninespine sticklebacksmature with 2 yr.

1049

TABLE l3.-Percent of Apostle Islands sticklebacks matureat each annulus with total number of fish in each groupin parentheses.

I's I's 50 mmunder 50 mm and over " III IV V

Male 34 56 93 100(56) (40) (78) (10)

Female 5 35 97 100 100 100(89) (67) (106) (44) (15) (2)

Within the ovaries, several developmentalgroups of ova can be distinguished, The ovarieswere classified into the following stages formaturity:

Immature (I)

Preserved in Formalin, the ova was translu­cent, spherical, and invisible to the naked eye.Ova of this group are present in the ovariesthroughout the year. Ovaries in this stage areslender, elongate, white organs.

Intermediate (II)

The first yolk granules appear in this stagegiving the ovary a pale yellowish tinge whilemaking the ova visible without magnification.

Maturing (Ill)

The yolk granules appear as highly refractive,spherical bodies in the cytoplasm, and ova be­come opaque. The ovary becomes turgid anddeep yellow in color, with ova readily visible butfirmly embedded in the follicles. The size of ovain this stage varies widely.

Ripe (IV)

When the ova are about ready to be spawned,a single conspicuous golden oil globule appearsin the yolk. The ova readily break from the fol­licles in which they were formed and flow freelyfrom the oviduct when the sides of the fish arepressed. In Formalin-preserved fish, mature ovacan readily be separated from ovarian tissuewith a dissecting needle, a procedure which wasalmost impossible in Stage III fish. The fish isvery rotund.

1050

FISHERY BULLETIN: VOL. 71. NO.4

Partly spent (V)

Partly spent fish contained ripe ova but notthe full complement. The ova are congregated inthe anterior ends of the ovaries, with the poste­rior portion somewhat flaccid and lighter incolor.

Spent (VI)

Ovaries were gradually resuming the appear­ance of those in the intermediate stage. Ovamaturing for the next reproductive period arealready visible to the naked eye. Followingspawning, resorbing ova remnants were some­times found.

Female fish were judged to be sexually maturewhen their ovaries contained ova in' develop­mental Stage II or beyond. Male maturity wasestablished by noting the presence of breedingcolor, an ink-black patch on the ventral surfaceas noted by McKenzie and Keenleyside (1970),remnants of which were detectable through theyear after its original appearance.

Time of Spawning

The first spent females were observed on June6 with no great increase in percentage frequencyof occurrence until mid-July. By the end of July,all but one of 38 fish were spent, and that onewas partly spent. It appears that the spawningseason for these fish is approximately 8 wk long.Equally long spawning periods are reported byNelson (1968b) and Bigelow and Schroeder(1953).

Spawning Habitat

The spawning act was not observed in thelake, therefore the exact location and spawningsubstrate cannot be identified. McKenzie andKeenleyside (1970) saw Lake Huron PlIlIgitillsspawning among rocks around the shore. Nosuch activity was noted in superficial observa­tion in the Apostle Islands. Ripe station 3 fe­males and males displaying vivid reproductivecoloration were found exclusively in the densegrowth of Nitl'l/u sp. At station 2, ripe fish ofboth sexes were collected over highly organic

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

where Y = number of eggs andX = length of fiRh in millimeters.

TABLE IS.-Comparison of mean number of eggs perfemale in fish taken in early June 1970 ahout 3 wk primto spawning and fish which were in "running-ripe"condition.

Fish length "Green" "Running-ripe"(mm) flsh fish

58 65.00596061 82.0062 88.50 61.0063 99.00 72.7564 98.66 68.0065 99.62 76.5666 99.33 72.4567 116.80 76.5768 114.50 78.8269 114.66 77.6170 114.00 87.2471 124.25 77.8772 137.75 95.7573 126.00 94.7074 130.66 92.4975 128.00 87.2876 132.00 111.3277 112.3078 158.00 95.7079 103.0080 103.0081 98.00

Total fish 58 132

"green" fish were significantly greater (0.01level) than those in corresponding size of ripefish (Fl,29 = 7.67). It appears atresia occurs inmaturing ninespine stickleback eggs just priorto spawning. Eggs which appeared to be degen­erating after some maturation were almost al­ways present in maturing ovaries. The decreasein egg number at each length interval presentedin Table 15 ranged from 24 to 39%. The amount ofatresia was not dependent on fish size. Vladykov(1956) estimated atresia affected up to 60% ofmaturing eggR in natural brook trout popula­tions. It is generally agreed that atresia is aphysical reaction to the pressure caused by thegrowing ovary (Vladykov, 1956).

When means of ripe egg numbers at eachmillimeter body length interval were plotted forRtations 2 and 3 nineRpine sticklebacks, linearrelationships were implied (Figure 8). Leastsquares equations which deRcribe the lineR were:

y = ~ 79.75 + 2.37XY = - 67.81 + 2.12X

Station 2Station 3

55 0 84.0 1.217 156 0 057 0 058 0 65.0 1.407 159 0 75.0 1.453 160 0 061 0 062 0 61.0 1.662 163 86.0 1.844 2 59.5 1.737 264 0 68.0 1.712 165 74.3 1.827 3 77.7 2.013 b66 76.0 1.896 3 69.8 1.901 467 78.8 1.997 10 71.0 2.101 468 82.7 2.031 6 73.0 2.206 469 76.8 1.961 5 78.3 2.288 670 86.6 2.142 5 87.7 2.371 771 82.8 2.254 4 71.3 2.544 372 93.4 2.429 7 104.0 2.452 273 103.2 2.639 5 82.6 2.766 574 98.3 2.400 3 90.0 2.730 775 88.0 2.742 5 85.5 2.920 276 113.8 2.942 7 94.0 2.776 177 112.3 2.990 3 078 95.7 3.070 3 079 0 103.0 2.962 180 103.0 3.010 1 081 0 98.0 3.522 1

Station 2 Station 3

Mean Mean Number Mean Mean NumberLength number flsh of number flsh of(mm) eggs weight fish eggs weight fish

(g) (g)

TABLE l4.-Average number of eggs counted from 132ninespine sticklebacks of various lengths collected 27June 1969.

muds 12 to 24 fathoms deep, and Lake SuperiorPU1Igitiu8 spawned successfully in this mud inthe laboratory while no spawning resulted inrocky or sandy substrates (Griswold and Smith,1972).

Fecundity and Egg Size

On 27 June 1969, stickleback samples at sta­tion 2 and 3 contained a large proportion offemales which were in Stage IV of maturity.These fish were used in the analysis of fecun­dity. A total of 72 fish were examined from sta­tion 2 and 60 from station 3 (Table 14).

Several authors have found a high percentageof atresia in maturing eggs of a number of spe­cies. This may cause an overestimate of fecundity(Wagner and Cooper, 1963; Vladykov, 1956;Wydoskii and Cooper, 1966). To determine ifappreciable atresia occurred in sticklebackeggs, the number of eggs in the 132 ripe fishwere compared to egg numbers in fish takenfrom the same stations about 3 wk prior tospawning (Table 15). Analysis of covariance(Snedecor, 1956) revealed egg numbers in the

1051

FISHERY BULLETIN: VOL. 71, NO.4

130r------------------,

'20

a:: t 10w~ 100=>z 90

~ 80w

70

60

63 65 67 69 71 73 75 77 79 81

FISH LENGTH (mm)

FIGURE 8.- Least squares regression lines indicating therelationship between fish length and number of eggs.Solid line and dots, station 2; dashed line and open circles,station 3.

Mean weight of ovaries containing ripe eggswas 10.2% and 9.8% of mean fish weight atstations 2 and 3, respectively. Since ovariesin immature fish are very small and somesomatic growth had already taken place byspawning time, it is clear that gonad matura­tion requires a gain in weight of over 10% ofthe winter weight of the fish. However, earlysomatic growth in mature females is consider­ably less than in males (Table 7). It appearsthat most early annual weight gain in femalesgoes into egg production and rapid somaticgrowth does not occur until after spawning.

Sex Ratio and Sex Specific Mortality

IOj4 (7c(n.68 (4dJ)

22.190dJ)**2()j70dJ)**

0.692: 1.714: 1.366: I.197: I.211: 1.137: 1.100: I

M"/t':Fema/e

0.930: 1.809: 1.339: I.12): 1

all female

Mom"

AprilMayJuneJuly

AugustNovemberDecember

oIIIIIIIV

The sex ratio of sticklebacks was determinedfor various years of life and for various monthsbecause it was obvious from general observa­tion that females predominated in the samples.All 808 fish from which otoliths were extractedwere used in the sex ratio calculations. Theratios were:

2The X values for age-groups II and olderwere significant at the 0.01 level, indicatingthe ratios were significantly different from 1to 1.

The sex ratios by month of all fish 1 + andolder beginning in April were:

Monthly length-frequency histograms (Figures5-6) indicate the presence of two distinct age­groups in most months, the older fish beingstrung out in an uneven pattern beyond the

Coefficient of determination (]'2) values were0,71 and 0.72 for stations 2 and 3 respectively.

Differences in fecundity relationships be­tween the two stations were implied and an­alysis of covariance revealed a significant dif­ference (0.01 level) in elevation of the two re­gressions (F j 28 = 19.0). With the presence ofatresia, the most obvious theory for this differ­ence would be a slight difference in time ofspawning at the two stations. This is suggestedby the slightly smaller mean weights of femalefish from station 2 sticklebacks at the time ofthese ovary collections (Table 14). This couldbe interpreted to mean station 2 females hadnot matured as much as station 3 fish sincefemale sticklebacks increase their weight byas much as 10% during the maturation period.Atresia would then have acted on the station2 fish to reduce their fecundity. Any differ­ence in spawning time is probably related toenvironmental differences between the sta­tions, exemplified by the presence of rootedaquatics at station 3 and their apparent ab­sence at station 2.

Measurements of egg size of six P. pungi­tillS, 62-79 mm long, were made on eggs de­posited in laboratory aquaria, since Vrat (1949)found stickleback eggs reached maximum sizeonly on extrusion into the water. Eggs wereleft in the water about 15 min before theywere removed and measured under a binocularmicroscope equipped with an optical microm­eter. Egg size appeared to be rather constantwith no relationship to size of fish. Egg diame­ter ranged from 1.53 to 1.98 mm (r = 1.76 mm).

1052

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

second mode. However, the second mode dis­appeared in August as young-of-the-year be­gan to appear in the catches. Since the 2-yr­old fish, which remained after August were pre­dominantly female (Table 16), it appeared thedie-off among old fish, as indicated by length­frequency diagrams, was more prevalentamong males. The monthly sex ratios of theage-groups comprising most of the spawn­ing population seemed to indicate that suchmortality began to occur as early as June witha sharp drop in males in the catchable popu­lation through July. However, the die-off didnot appear in the length-frequency diagramsuntil August. This discrepancy might be ex­plained by a differential movement of malesonto the spawning grounds in June to set uptheir territories and/or an increased aware­ness of the fishing gear by territorial males.Both these explanations could result in de­creased male availability to the trawl duringthe spawning period. However, after the spawn­ing period, males remained low in the catchthroughout the rest of the year, indicating ahigh mortality among males which spawned.

These interpretations are strengthened bytwo other observations. Males which hadspawned previously were identifiable at anytime by vestiges of spawning colors on theirventral surface. Few of these individuals wereseen in the collections. Also, in a laboratoryexperiment described by Griswold and Smith(1972), eight males brought off broods to thePoint where fry were actively swimming aboutin the aquaria. Seven of these males died withina month of spawning. There was no mortalityobserved among 10 females which spawned inthe laboratory.

TABLE 16.-Ratio of male sticklebacks to females ofdifferent age-groups in various months. Annulus formationis complete after July.

Age Apr. May June July Aug. Nov. Dec. Mean l

0 0.5 1.2 0.7 2.0 0.9 0.5 0.91 .7 1.1 .2 0.3 .6 .9 1.6 .82 .3 .3 0 .1 .2 .4 .4 .33 0 0 0 .1 0 .1 .3 .14 0 0 0 0

Mean2 .4 .7 .3 .2 .7 .7 .7

~ Weighted for number of fish in each month.Weighted for number of fish in each age-group.

PARASITE INFECTION

The pleurocercoid stage of the cestode, Schi­stocephalus pUllgitii Dubinina (Pavlovskii,1964), is the only parasite generally abundantamong ninespine sticklebacks. The procercoidinfects numerous species of copepods, and whenthese are eaten by sticklebacks, the procercoidburrows through the stomach wall to grow as apleurocercoid in the body cavity. At this stagethe parasite is species-specific as, so far as isknown, it attacks only the ninespine stickle­back. Two months arp required for full growthof the pleurocercoid, at which time it may in­fect a number of fish-eating birds. No otherstickleback parasites were noted in the study.

Rate of infection in Apostle Islands stickle­backs was low in 621 fish examined. Only 4 of362 fish (1.10%) were infected prior to 1 Augustin all years of the study combined, and 13 of259 fish (5.01%) were infected after 1 August.Summer and fall infection would intuitively behigher because of greater copepod abundance atthat time. The largest pleurocercoid found was22 mm long and weighed 0.225 g or 14% of thehost's weight. Two fish had two pleurocercoidseach which made up 18 and 14.5% of the fish'sweight, respectively. These parasites were foundin 7 male and 10 female sticklebacks. Fish in­fected at spawning time did not appear to bematuring.

FOOD HABITS OF THE STICKLEBACKAND ASSOCIATED SPECIES

Five papers mention the food of P. pllllgdius.Forbes (1883) found two specimens which con­tained equal amounts of diptera larvae andcladocerans. Blegvad (1917) found copepodsand stickleback eggs predominant in 112 Danishfish. Hynes (1950) reported cladocera, amphi­pods, and chironomid larvae were the mostimportant food items in fish from an Englishstream. Bigelow and Schroeder (1953) state thatthe stickleback feeds voraciously on fish eggsand fry and can be very destructive to otherspecies in fresh water. Anderson and Smith(1971) report crustaceans make up more than90% of the Lake Superior stickleback diet. In thepresent study, emphasis was placed on the food

1053

hahits of sticklebacks and several associatedspecies with which they are most likely to com­pete or which may be involved in predator-preyrelationships with sticklebacks.

In the stomach analyses, contents were re­moved from the upper esophagus to pyloricsphincter and measured by water displacementin 15-cm3 centrifuge tubes. The contents werethen emptied into a gridded petri dish for indenti­fication under a binocular microscope at 10 xmagnification. Organisms present in. smallnumbers were counted, but larger numbers wereestimated from counts in randomly selectedsquares of the dish. Percentage of each item inthe total volume of each stomach was estimated.The data are reported as percentage frequencyof occurrence and percentage of total volume foreach food item in stomachs containing food.

Invertebrate Abundance

Invertebrates important in the sticklebackdiet were sampled with the mysis sled concur­rently with sticklebacks used in food analysis;therefore, it is assumed that the gear sampledfood organisms that were available to stickle­backs. One-mile tows were made with the sledand the invertebrate catch was preserved in10% Formalin. Invertebrate samples were identi­fied and enumerated on a gridded petri dish inthe same manner as stomach contents.

The average number of organisms per 100 mof tow is shown in Table 17 for 18 tows on fivedates. The data indicate the extreme abundance

FISHERY BULLETIN: VOL. 71. NO.4

of POlltopo)'eia affilli8 adjacent to the bottom.Although copepods were also abundant, theywere usually not as important in mass as eitherPOlltopo)'eia or M!J8i8. Coregonid eggs and frywere found occasionally in fall plankton samples.

Food of the Stickleback

Food analyses of 421 stickleback stomachs innine different months were made (Table 18).

Fish under 50 mm long fed primarily onFOlltoporeia and copepods, which made up 79and 13% of the total food volume, respectively.Cladocerans were of minor importance, andmysids were also present infrequently. Otherorganisms present in these smaller fish werechironomid and diptera larvae and a filamen­tous green alga.

Food of sticklebacks 50 mm and over werecharacterized by a variety of larger food items,suggesting that one factor limiting the diet ofyoung fish is size of food items. FOil toporf'iataken by the under 50-mm sticklebacks weresmall in size (4.0 mm average) compared withthose found in adult stomachs (6.5 mm).

In these larger fish, FOllto]Joreia made up61% of the total stomach volume, MY8i8 21%,copepods 8%, and miscellaneous other items10%. FOil topm'cia was the major food item in allmonths but November. There is a suggestionfrom late fall and early spring samples that core­gonid eggs may be an important food item inthe winter. Except for these eggs, there is littleto suggest that F. JlllllrJitill8 is significantly

TABLE 17.-Average number of invertebrates caught per 100 m in tows along bottom with mysis sled in the Apostle Islands.

[The percentage volume of each item found in stickleback stomachs is in parentheses.1

Organism 29 Apr. 1968 22 July 1968 13 Nov. 1968 29 Apr. 1969 27 June 1969

Number of tows 4 3 4 2 5

Amphipodo,Pontoporeia llffini\' 437,750 (88) 748,260 (61) 117,600 (33) 32,360 (84) 371 ,660 (69)

Mysidocea,Mr.\'is relicta 120 (4) 6,110 (12) 34,280 (60) 1,370 (8) 320 (9)

Cladocera 3,170 (1) 1,040 (4) 280 (0) 30 (2) 2,610(3)Bosminll spp. 2,210 910 190 30 2,330Daphnia spp. 920 120 90 0 90Others 40 10 0 0 140

Fish eggs 0(0) 0(0) 0(0) 0(6) 0(0)Capepods 43,460 (7) 98,630 (23) 47,210 (0) 23,500 (0) 199,790 (16)Diptero larvae 0(0) 0(0) 0(6) 0(0) 0(3)Mollusca

Sphaeri idee 2 (0) 0(0) 0(0) 1 (0) 0(0)Other 0(0) 0(0) O(l) 0(0) 0(0)

1054

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

TABLE IS.-Food of Apostle Islands sticklebacks expressed as percentage frequency of occurrence and (in parentheses)percentage of total volume for each food item.

Sticklebacklength Item Apr. May June July Aug. Sept. Oct. Nov. Dec.

Under 50 mm Number of Stomachs 0 0 13 18 4 0 5 14 0Number empty 0 0 2 1 0 0 0 0 0

Crustaceans:Pontoporeia affinis 62 (66) 100 (99) 100 (100) 93 (86)Copepoda 15 (11) 100 (97) 7 (3)Cladocero 20 (3)Mysis relict a 6(1) 7 (7)

Insects:Chironomidae 8 (9) 14 (4)Diptera

Algae 8 (9)

50mm Number of stomachs 31 66 99 27 27 55 20 20 22and over Number empty 11 4 3 3 0 1 1 0 2

Crustaceans:82 (62)P01Jl()poreia «[finis 80 (75) 61 (49) 90 (75) 30 (24) 92 (92) 90 (80) 60 (30) 59 (45)

HJ'aUe/a sp 1 (1)Ml'sis rtJliC1U 5 (1) 56 (42) 30 (8) 11 (10) 44 (30) 25 (1) 70 (65) 36 (32)C';pepoda 5 (2) 12 (3) 8 (13) 52 (38) 4 (2) 5(1) 5 (12)Cladocera 5 (2) 1 (1) 33 (9) 11 (6) 2(1) 5 (2)

Insects:Chironomidae 9(1) 11 (5) 16 (3)Other Diptera 2(1) 7 (6) 4 (1)Ephemeroptera 1 (1) 4 (3)

Nemotoda 2(1)

Coregoni d eggs 5 (2) 12 (3) 27 (20)

Unidentified 5 (15) 1 (2) 3 (2) 7 (7) 22 (3) 5 (5) 5 (5)

destructive to fish eggs and fry in Lake Superioras Bigelow and Schroeder (1953) thought wasgenerally the case. Neither eggs nor fry of anyother kind were found in stickleback stomachs,although gravid females and young-of-the-yearof smelt, trout-perch, slimy sculpins. and pygmywhitefish were taken in the same areas at aboutthe same time as the stickleback stomach sam­ples.

The percentage volume of total stomach con­tents for each day and for each food item takenin the mysis sled is given in Table 17. Po/!fo­pore/a was the most important food item exceptin the November sample. In that month. mysidswere the most important food and were moreabundant in plankton samples than at any othertime. Copepods were most abundant in the Juneand July samples, and their relative contribu­tion to stickleback diet was also highest in thesemonths. In general, when mysids or copepodsbecame relatively more abundant with respectto themselves or to other zooplankters in the

plankton samples, they also become more im­portant in the stickleback diet. This suggeststhat the stickleback shows no strong food pref­erences but is adaptable and eats what is mostavailable.

Relation to Food Utilization ofOther Forage Species

Juvenile smelt (less than 12.5 em), trout­perch, and slimy sculpins were found to havediets similar to sticklebacks in Lake Superior(Anderson and Smith, 1971). Since all thesespecies were common in the Apostle Islands,competitive relationships between them and thestickleback could exist. A rank correlation coef­ficient was used to measure the similarity of foodcomposition between two samples and thuspermit examination of interspecific and intra­specific feeding relationships.

Table 19 is an example of the computationas discussed by Snedecor (1956). In the case of

1055

TABLE 19.-An example of the cakulation of rank­correlation coefficients.

Food item 10 stickleback 10 trout- Differencecategory rank perch rank d d 2

p()nloporeia 1 1 0 0M.rsis 3 4 -1 1Copepods 2 3 -1 1Chironamidae 5 2 +3 9Cladocera 4 5 -1 1

" 5 Ed = 0 'I:d" = 12

1 - 6(12)= 1 72rs = 5(25 -1) 120 = 0.40

a tie in importance of any two categories, thelarger Item was ranked first. An r.. value of+ 1.0 indicates that the various food categorieswere eaten in the same order of preference, andas rs approaches -1.0, there is no agreementin order of occurrence.

To test the hypothesis that fish of the samespecies eat the same things in different areas,10 randomly selected fish caught from one sta­tion within 3 days were compared. In interspe­cific tests, 10 fish of each species taken at thesame station on the same day were used. Except

FISHERY BULLETIN: VOL. 71. NO.4

for smelt, only adults were used in the testsbecause juveniles of other species were not com­pletely available to the gear and not taken inlarge enough quantities to test.

Analyses of individual tests were not made;instead the results from several comparisons(N in Table 20) were combined and a test madeusing the combination of probabilities discussedby Snedecor (1956:216). The sum of the naturallogarithms of the probabilities associated witheach of several comparisons is multiplied by- 2, and the resulting values have a chi-squaredistribution with two degrees of freedom foreach comparison. Pearson (1924) presents afigure which allows graphical calculation ofprobabilities associated with r,.

Intraspecific correlations were significantfor all comparisons (Table 20). Indications arethat food habits among individuals of the samespecies were almost identical within the ApostleIslands.

Significant interspecific correlations werefound between sticklebacks and slimy sculpinsin spring and summer and sticklebacks and

TABLE 20.- Food composition similarity measured by rank-correlation coefficients (rs) between stickle­backs and other Apostle Islands species. (N equals the number of individual tests for each comparison,and n equals the number of food categories in each comparison.)

Comparison

1. Between adult sticklebacksof stations 2 and 3

2. Between adult trout·perchfrom two stations

3. Between juvenile omeltfrom two stations

4. Between adult sculpins fromtwo stations

N " Mean rs Chi-square

5 6 +0.94 36.89**

3 5 +.83 16.81*

3 4 +.85 16.98*

5 6 + 1.00

5. Between adult sticklebacksand other species withineach catch:

Spring:Trout -perchSmeltSculpins

Summer:Trout -perchSmeltSculpins

Fall:Trout -perchSmeltSculpins

7 5 + .70 30.89*10 4 + .40 34.2910 6 +.89 62.02**

4 5 + .43 13.942 4 - .30 6.54

10 6 + .89 62.02**

4 5 + .33 13.284 4 +.40 13.725 6 + .37 16.87

* Significant at the 0.0 1 level.** Significant at the 0.005 level.

1056

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

FIGURE 9.-Percent volume of various food items found inlake trout 28 em and longer at various sampling stationsin the Apostle Islands. Fish were taken in June and Sep·tember during the two annual BeF index trawling cruises.

in.) long (TL). Only 3.5% of the volume of stom­achs from 142 fish under 25.4 cm (10 in.) wassticklebacks. In 294 trout from 25.4 to 28.0 cm,sticklebacks made up 32% of the volume of thediet.

Sticklebacks were a major part of the diet oflake trout from 28.0 to 38.0 cm (11.0 to 14.9 in.)(Figure 9). P. pUligitius made up 39% of volumein the diet and were found in 42% of the 176trout which contained food. Smelt was the secondmost abundant fish species in the diet of 28- to38-cm trout. The food habits of trout remainedabout the same for all ice-free months (Andersonand Smith, 1971).

The major food of trout longer than 38 cmwas smelt. Of 30 trout in this size range, 69% ofthe stomach contents were smelt remains.Sticklebacks were found in five of these fish andmade up 17% of the volume. Nineteen adultsticklebacks and two sculpins were found in one50-cm (19.7 in.) trout. Sticklebacks eaten bytrout were all age-group I or older. No stickle­back eggs or age-group 0 fish were found inover 4,500 trout stomachs sampled on the LakeSuperior project.

The explanation of the absence of small stick­lebacks in the trout diet is probably one of dis-

trout-perch in spring. The significant rank cor­relation does not necessarily indicate theamount of competition for food, but merely thepotential for such competition. For competitionto occur, there must be a demand by more thanone organism for the same food in excess ofimmediate supply. It can only be said that sincethe abundance of sticklebacks and slimy sculpinsare positively correlated in the index catchesand their food relationships are similar, thepossibility of competition between the two spe­cies exists during some seasons. The actual ex­istence of competition is highly questionablebecause of the apparent abundance of POllto­poreia, the number one food item, and the ap­parent adaptability of the stickleback in foodutilization. The drop in correlation in the fallwas due to greater use of chironomids bysculpins.

While the correlation with trout-perch alsoindicates potential competition for food, theirabundance in the index trawl samples was nega­tively correlated and Dryer (1966) foundtrout-perch and sticklebacks to have differentbathymetric distributions. While the trout-perchand sticklebacks utilized the same foods inspring, their secondary foods were different insummer and fall. Anderson and Smith (1971)found trout-perch fed on wider variety of fooditems than sticklebacks in the Apostle Islands.This combined evidence seems to indicate thatcompetition between these two species is alsounlikely.

Sticklebacks as Forage for Other Fish

Several fish species which occur in LakeSuperior have been reported to feed on stickle­backs at various locations. These include laketrout; brown trout, Sa/1l1Ii trutta L.; rainbowtrout, Sa/mo gaird II eri Richardson; burbot,Lota Iota (L.); and various sculpins. Of these,the lake trout was the only species which fedheavily on sticklebacks in the Apostle Islands.

Dryer, Erkkila, and Tetzloff (1965) related achange in diet in juvenile Lake Superior laketrout to their size. The diet, primarily a crusta­cean-insectivorous type in small trout, becamepiscivorous in larger fish.

In the present study, sticklebacks were foundin 7.5% (33) of 436 lake trout less than 28 cm (11

80

w~:::l 60-J0>~zwU 40a:wa.

20

STA. 86 STA. 7~ ST". 44 STA.2 ST". 12

1057

tribution. After hatching, the young gticklebackscongregate in the shallow sandy areas wherethey are unavailable to lake trout and mostother species. By the time they leave the shal­low water in late fali, they have grown to 45 mmor more. It is at this length that they begin toappear in trout gtomachs the following spring.

Smaller trout cannot feed well on stickle­backs once the latter become available after asummer's growth, because of the protectivespines. This inability explains the relative un­importance of sticklebacks in the diet of smalltrout. Similarly, Hoogland, Morris, and Tinber­gen (1957) found young northern pike, Eso.l'

lucius L., rejected gticklebacks, first exhibitinga negative reaction to mechanical stimuli uponcontact with the gpineg and then a visual avoid­ance. Upon reaching a certain size, the pike pref­erentially take sticklebacks (Frost, 1954). Adiscussion by Fortunatova (1959) supports thistheory. The dietary shift to smelt by largertrout can be explained by assuming the optimumsize of the prey increases as the trout growlarger.

Compared to all lake trout over 28 em longtaken from a number of trout index gtations,fish from stations 75 and 86 ate a high percent-

FISHERY BULLETIN: YOLo 71. NO.4

age of sticklebacks (Table 21, Figure 9).At station 86, considered the best trout nur­sery area in the Apostle Islands, sticklebackswere the predominant food item even thoughjuvenile smelt were more abundant in the indextrawl samples (Table 9). At station 75, thesecond most important nursery area, stickle­backs were very abundant, and they made upmore than 60% of the trout diet. Station 12sticklebacks and smelt were about equal in theindex catch, but sticklebacks were much moreimportant in the trout diet. At station 44, thepredominant forage fish in the trawl sampleswas sculping, and these were dominant in troutstomachs. At this station, sticklebacks werecomparatively rare in occurrence, but gtillmade up 21% of the volume of the trout diet. Itmight be assumed from station 44 data thatgculpins are an important item in the diet andthat trout will substitute sculpins when stickle­backs and gmelt are not readily available. How­ever, the stations with high sculpin abundancerelative to gtickleback abundance do not supportlarge numbers of trout. From these data, itappears that the availability of sticklebacks atthe various gtations is probably important inthe gelection of these areas by lake trout as

TABLE 21.-Percent frequency of occurrence and (in parentheses) percent total volume for each fooditem found in lake trout 28 em and longer at various sampling stations in the Apostle Islands. Fish weretaken in June and September during the two annual Bureau of Commercial Fisheries index trawlingcruises.

Station Station Station Station StationItem 86 75 44 2 12

Number of stomachs 77 45 24 51 33

Number empty 20 3 0 0 0

Crustaceans:Mysidoceo 35 (25) 15 (4) 33 (22) 16 (6) 64 (29)Amphipodo 18 (4) 8 « 1) 17 (12) 2 « 1) 36 (10)Copepoda 1«1) 3 « 1)

Fish:Smelt 14 (18) 9 (8) 12 (16) 21 (25) 6 (6)Stickleback 30 (36) 64 (62) 17 (21) 21 (23) 33 (29)Sculpin 6 (7) 13 (13) 21 (25) 12 (10) 6 (4)Unidentified 9 (8) 15 (11) 17 (3) 27 (36) 24 (19)

Fish eggs 1 « 1)

Insects:Chironomidae 6 «1) 6«1) 2«1) 3 «1)Coleoptera 1«1)Trichoptera 1«1) 1 «1)Other 1«1) 7 « I) 1«1) 6 (3)

Miscellaneous 1«1)

1058

GRISWOLD and SMITH: LIFE HISTORY OF NINESPINE STICKLEBACK

nursery areas, although other factors are alsoimportant because there are many areas wheresticklebacks are abundant and trout are not.Conversely, no good trout nursery area areknown where sticklebacks are not abundant.

The importance of the stickleback as a fooditem for the lake trout in the Apostle Islandsappears to be substantial while the latter arebetween 28 and 38 cm long. No information isavailable, however, to indicate the amount ofsubstitution which would occur if the stickle­backs were less available.

LITERATURE CITEDANDERSON, E. D.

I \16\1. Factors affecting ahundance of lake herring,CO/"cgol1lis artetJii Lcsl'ur. in western Lake Superior.Ph.D. Thesis, Univ. Minnesota. Minneapolis, 316 p.

ANDERSON. E. D., AND L. L. SMITH, JR.

1\171. A synoptic study of food hahits of 30 fish speciesfrom western Lake Superior. Univ. Minn. Agric.Exp. Stn. Tech. Bull. 27\1, 1\1\1 p.

BAILEY, M. M.

1\164. Age, growth, maturity, and sex composition ofthe American smelt, 0,/1/('1'/1.\ IIIl1rd(/x (Mitchil1l,of Western I.ake Superior. Trans. Am. Fish. Soc.\lUH2-195.

BERTIN, I..\925. Recherches hionOll1ll\UCS, hiolnct ril\ucs d

systcillatillues sur les cpinoches ((;(/.,lhll.l/';/lI,;.Il.

Ann. Insl. Oceallogr. Monaco, Nouv. Ser. 2: 1-204.BIGELOW, H. B .. AND W. C. SCHROEDER.

1951. Fishes of the Gulf of Maine. U.S. Fish Wildl.Serv., Fish. Bull. 51, 577 p.

BLEGVAD, H.

1\117. On the food or fish in the Danish waters withinthe Skaw. Rep. Dan. BioI. Stn., Dep. Agric. 24:l'i- 72.

CARLANDER, K. D., AND I.. I.. SMITH. JR.

1\144. Some uses of nomographs in fish growthstudies. Copeia 1\144: 157-162.

DRYER, W. R.1\161. Age and growth of the whitelish in Lake

Superior. U.S. Fish Wildl. Serv .. Fish. Bull. 61: 77­\I'i.

I \166. Bathymetric distrihution of fish in the ApostleIslands region, Lake Superior. Trans. Am. Fish.Soc. \15: 24H-2'i9.

DRYER, W. R.. AND J. BEtL.

I %4. Life history of the lake herring in Lake Supe­rior. U.S. Fish Wildl. Serv .. Fish. Bull. 63:4\13-5:lO.

DRYER, W. R., I.. F. ERKKILA, ANll C. I.. TETZLOFF.

196'i. Food of lake trout in Lake Superior. Trans.Am. hsh.Soc.\l4:169-176.

FORBES, S. A.lXX3. The food of the smaller freshwater fishes. Bull.

Ill. Nat. Hist. Surv. 1(6):65-\14.

FORTUNATOVA, K. P.1959. Availahility of sticklebacks as food for the

predacious fishes of the Volga delta. (Translatedfrom Russ., 1\161, Fish. Fish. Res. Board Can.Transl. Ser. 331, Nanaimo, B.C.. IH p.)

FROST, W. E.1954. The food of the pike, E.HiX III('ill.l L.. in Winder­

mere. J. Anim. Eeul. 23:33\1-360.GRISWOLD, B., AND L. I.. SMITH, JR.

I \172. Early survival and growth of the ninespinesticklehack, PIII/gilill.l {Jllllgilill,'. Trans. Am. Fish.Soc. 10 U50-352.

HOOGLAND, R., D. MORRIS, AND N. TINBERGEN.

1956. The spines of sticklehacks (G(/.Item.l/l'lI.1 andPrgll,,'tl'llS) as a means of defense against predators(P('/'w and EHiX). Behaviour 10:205-236.

HUBBS, C. L.. AND K. F. I.AGLER.

195H. Fishes of the Great Lakes region. Revised cd.Cranbrook Inst. Sci. Bull. 26, 213 p.

HYNES, H. B. N.

1950. The food of fresh-water sticklebacks (Ga.l/('/'­

IISlell.l a('/<lealll.l and PI'guslell., Plil/gilill.l), with areview of methods uscd in studies of the food offishes. J. Anim. Eeul. 19:36-5H.

JONES. J. W., AND H. B. N. HYNES.

1950. The age and growth of G(/.I'tem.l'lell.l (/('1111'(/111.1,

p.l'guS!CllS piingilius and Spinachia \'ldgaris as shownhy their otoliths. J. Anim. Ecol. 1\1:59-73.

I.EE, R. M.

1912. An investigation into the methods or growthdetermination in fishes. Cons. Perm. Int. Explor.Mer. Puh\. Cireonstance, 63, 34 p.

LEtNER, M.1934. Die drei Europaichen sticklinge, (Ga.\'lem.,/l'II.,'

£lell/catlls L.. G. PUtlR;tius L., G. SpillllChi£l L.)und i1ue kreuZllngsprodukte. Z. Morph. Oekol.Tiere 2H: 107-154.

McKENZIE, J. A., AND M. H. A. KEENI.EYSIDE.

1970. Reproductive hehavior of ninespine stickle­hacks (Plil/gilill., plil/gilill.\' (1..» in South Bay,Manitoulin Island. Ontario. Can. J. Zoo!' 4R:55­61.

MOORE, H. H., AND R. A. BRA EM.

1\165. Distrihution of fishes in U.S. streams tributaryof Lake Superior. U.S. Fish Wildl. Serv., Spec.Sci. Rep. Fish. 5 16, 61 p.

NELSON, J. S.

1\16Ha. Deep-water ninespine sticklebacks, Ptll/gitill.\'

PilI/gil ill.', in the Mississippi drainage, CrookedLake, Indiana. Copeia 1\l6H:326-334.

1\l6Rb. Ecology of the southernmost sympatricpopulations of the hrook sticklehack, Clllae(/ il/­(,ol/.\'Ial/.\', and the ninespinl' sticklehack, Plil/gil;ll.\'

pllI/gilill,', in Crooked Lake, Indiana. Proc.Indiana Aead. Sci. 77: lR5-1\12.

PAVlOVSKlI, E. N. (editor),\964. Key 10 parasites of freshwater fish of the USSR.

Izd. Akad. Nauk SSSR, Moscow, \I 19 p. (Translatedhy Israel Program Sci. Transl., 1%4; availahleU.S. Dep. Commel., Natl. Tech. InL Serv.,

Springfield, Va., as TT 64-11040.)

l05H

PEARSON, K.1924. Tables for staltstlClans and biometricians. 2d

ed. Cambridge Univ. Press, 146 p.SNEDECOR, G. W.

1956. Statistical methods. 5th ed. Iowa State Univ.Press, Ames, 534 p.

VLADYKOV, V. D.1956. Fecundity of wild speckled trout (Salvelil/lls

fOI//il/ali.l) in Quebec lakes. J. Fish. Res. BoardCan. 13: 799-841.

VRAT, V.

1949. Reproductive behavior and development ofeggs of the three-spined stickleback (Gas(eYOS(ell.1

aClilea/II.I) of California. Copeia 1949:252-260.

1960

FISHERY BULLETIN: VOL. 71, NO.4

WAGNER, C C, AND E. L. COOPER.

1963. Population density, growth, and fecundity ofthe creek chubsucker, Erillly;:.on oblonKlis. Copeia1963:350-357.

WALFORD, L. A.1946. A new graphic method of describing the

growth of animals. BioI. Bull. (Woods Hole) 90:141- 147.

WVDOSKI, R. S., AND E. L. COOPER.

1966. Maturation and fecundity of brook trout frominfertile streams. J. Fish. Res. Board Can. 23:623­649.